Study: Model Hallucinations Can Be Linearly Decoded
ziv_ravid · x · 2026-07-04
Researchers found that a model's state of "running on imagined content" might be linearly decodable, meaning it can theoretically be detected. The author notes that grounding without imagination is a camera, imagination without grounding is a mirage, and the truly interesting part lies at the seam between the two.
More from Research
- LightMTP: lightweight latent multi-token prediction boosts LLM planning in pretraining — gdm3000 · 2026-10-27
- EMNLP study: AI literary translation reads fine but varies more, even within one book — tuvllms · 2026-10-27
- Lab monkeys at $30-80K each: a 4-year breeding cycle is the real preclinical drug bottleneck — zebird0 · 2026-10-27
- OpenAI Foundation funds UNC tumor antigen project, hiring senior experimental scientist — iskander · 2026-10-27
- Steering attention heads toward honesty: up moral patienthood, down psychopathy — edelwax · 2026-10-27
- Are transcoders overkill? New paper argues plain neurons suffice for circuit analysis — burny_tech · 2026-10-27